electronicmusicfandomcom-20200213-history
S770
A sampler produced by Roland from 1980 through the mid-1990s. For most of this time, it was Roland's top-of-the-line sampler. Complex and lavishly featured, it was used extensively by Hollywood soundtrack composers and foley artists, who valued its sonic fidelity and versatility. Except for the analog inputs and outputs, the circuitry was entirely digital; a set of DSPs which did the real-time processing were managed by an Intel 80186 CPU, which managed the interfaces and did offline processing. The S-770 was one of the first samples to use a full 16-bit width for its samplers, and sample at a high enough rate to capture a full 20 Hz - 20 KHz bandwidth, matching the specifications of the Compact Disk audio format. It was capable of mono and stereo sampling, and could store samples in its built-in disk drive or another drive connected to an external SCSI disk port, as well as on 3.5" floppy. It had the ability to import and do sample rate conversion on samples from several older Roland models. The hardware was lavishly equipped with I/O and user interface capability, the most notable of which was its interface for connecting an external video monitor. Although the unit contained an LCD display which was easily readable and large by the standards of that era, the OS and user screens and menus were so complex that the unit was best used and programmed by connecting an external color monitor; the LCD screen was adequate for on-stage performance as long as no patch editing or programming needed to be done. The external monitor connection used the now-obsolete CGA standard -- an unfortunate choice by Roland, since that standard was already becoming obsolete by the time the S-770 went into production. It is wired to connect to a monitor with an old-standard 12-pin rectangular connector. A special cable was produced by Roland for connecting to monitors with the 15-pin D-connector; it is no longer available, but fabricating one is not too difficult. The S-770 included a computer mouse for user input, in addition to function buttons under the LCD display, a few dedicated function buttons, and a jog wheel on the panel. The mouse conformed to the Japanese MSX standard, rather than the IBM AT standard used in the U.S. and Europe. Roland offered two additional input device options. The RC-100 remote control device was on the end of a long cord, which plugged into the sampler's mouse port. It duplicated the panel controls and included a few extras, such as a numeric keypad that could be used to enter exact sample word values. It also included a port for the mouse to plug into. Combined with a monitor on a long cable, this allowed all of the user interface to be remoted to, say, a mixing desk location, while the sampler itself remained in a rack or equipment closet. Roland also offered a pen digitizing tablet, which could be used as a substitute for the mouse. Audio I/O consisted of a pair of analog inputs for sampling, which were on the front panel for easy connection of microphones or other sources; these had a trim control for use with mics or line-level inputs, and a clipping indicator. The rear panel offered a stereo pair of main analog outputs, eight individual analog outs to which specific samples in a set could be routed, and a S/PDIF stereo output. Also on the back panel were the CGA video output, a composite NTSC video output (which produced only a monochrome image), and a SCSI port for connecting external disk drives; up to six could be connected. A fan was mounted on the back panel; this could be turned off for brief periods via a system menu, but it could not be left off without external cooling because the electronics would overheat. As mentioned, the feature-full operating system was complex and required study of the manual and practice in order to master. (It was sufficiently complicated that Roland, contrary to its usual practice at the time, hired an American to write a separate English-language version of the manual, rather than do a translation from the Japanese version.) It divided all of the information and parameters into four layers, using these names: * Volume, which... * Performance, which consisted of a collection of patches, each of which was assigned to a MIDI channel number, and could also be transposed and restricted to a specific range of the keyboard independently of the key split information programmed into the partial. * Patch, which... * Partial, which... * Sample, which contained the actual sample data, as well as some basic information about the sample, such as what key it was in. Each layer had a corresponding edit screen at which parameters at that layer could be edited. The edit screens include a "drill down" feature which allowed editing at the next lower layer. For instance, while editing a Patch, one could choose one of the Partials in the patch and select "edit partial from patch", and access that Partial for editing. When done, the user could then "pop" back to the Patch layer and resume where they left off at that laer. It was intended that for a performance, the performer would choose a screen at the Performance or Volume layer. At lower layers, the MIDI interface usually worked in omni mode, to remove one level of complication from the editing process. The voice architecture was the digital equivalent of the conventional VCO-VCF-VCA subtractive synthesis chain. The equivalent of the VCO was the sample playback oscillator, which allowed the user to set any of several loop points and start/stop points with use in different loop modes. Some of the loop modes allowed for different segments of the sample to be played during different envelope generator phases. The filter algorithms were similar to the ones used in the JD-800; they would not self-oscillate but were capable of being pushed to extreme values of resonance. They supported all of high pass, bandpass, and low pass modes. Not finished... more later...